Homelessness and alcohol dependence are commonly co-occurring and serious public health issues. Unfortunately, abstinence-based alcohol treatment approaches are minimally effective in engaging and successfully treating homeless individuals with alcohol dependence. There have therefore been calls for more flexible and client-centered approaches tailored to this population's needs. Innovative, low-barrier approaches (e.g., Housing First and alcohol management programs) have been applied with this population and are efficacious in reducing alcohol use and related problems as well as utilization of publicly funded services and associated costs. Such approaches have been referred to as harm-reduction interventions because they focus on reducing alcohol-related harm for affected individuals and their communities without requiring a commitment to abstinence-based goals. Although psychosocial, harm-reduction approaches are beginning to proliferate for this population, there are few pharmacological counterparts to support and enhance these efforts. One medication that could address this treatment gap is extended-release naltrexone (XR-NTX; marketed as Vivitrol). XR-NTX is a 30-day, extended release formulation of the opioid receptor antagonist, naltrexone, and is administered monthly via gluteal intramuscular injection. The proposed Phase II study features a four- arm RCT (N=300) designed to test the efficacy of XR-NTX as a pharmacological adjunct to existing psychosocial harm-reduction services provided by community agencies to homeless people with alcohol dependence. The proposed study will include a 24-week follow-up and will test the relative efficacy of 3 active treatment combinations-1) XR-NTX+harm reduction counseling, 2) placebo+harm reduction counseling and 3) harm reduction counseling only (HRC)-compared to the services as usual (TAU) that all participants receive from community agencies. This proposed design will allow us to dismantle active treatment components and thereby detect potential placebo effects of both the administration of an injection and attention from a medical professional. In this study, there are three primary specific aims. First, we will test the relativ efficacy of XR-NTX, placebo and HRC compared to TAU in decreasing alcohol quantity, frequency and alcohol-related problems. Second, we will test hypothesized mediators of the intervention effects. Specifically, we hypothesize that the active treatments will precipitate increases in motivation to change and decreases in craving, which, in turn, will mediate the active treatment effects on alcohol outcomes. Finally, we will test treatment effects on publicly funded service costs (i.e., emergency medical services, ER visits, hospital admissions, and county jail). It is hypothesized that XR-NTX, placebo and HRC groups will show greater decreases in publicly funded service costs than the TAU group.